Educational Design for Scaffolded, Interactive Simulation-Based Learning: SeaVarinv, an Open-Source GUI for Paleo-Sea-Level Reconstruction in Geoscience Education
Year: 2026
Type: article
Abstract: Sequence stratigraphy asks students to reason in both directions along the model chain: forward, from sea level to strata, and backward, from strata to sea level. The backward direction is an inverse problem—ill-posed, non-unique, and dependent on regularization choices that textbooks describe but rarely let students feel. Research codes that close this loop assume command line fluency and stochastic sampler expertise beyond classroom reach, while existing educational tools are passive or forward only. This paper presents SeaVarinv, an open-source, GUI-driven stratigraphic simulation tool, designed to make the full forward–inverse workflow accessible, interactive, and instructive for students without a programming background. Ten guided teaching modules, each mapped to the revised Bloom’s taxonomy, walk students from the forward map through non-uniqueness, noise sensitivity, regularization trade-offs, and uncertainty. Quantitative demonstrations ship as reproducible reference results: inversion recovers synthetic sea-level histories with RMSE values of 7.2–7.8 m and correlations of 0.92–0.94 across 1–5% noise; the L-curve knee at λ = 10−4 nearly matches the unknowable oracle weight; total variation preserves an abrupt 25 m step (RMSE 3.35 m) where Tikhonov smoothing fails (7.41 m); and a nominal 95% credible band achieving only 25% coverage supplies a concrete calibration failure lesson. All experiments are seed fixed and checkpointed. By making the complete forward-to-inverse loop tangible in a single lab session, SeaVarinv turns the hardest abstractions of computational stratigraphy into measurable, hands-on experiences.
Source: Applied Sciences
DOI: 10.3390/app162211238
Authors: Qiu Shi, Li Hua, Liu Yunuo, Dai Yang, Li Jingzhe
Language: English
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
Keywords: sequence stratigraphy, paleosea-level reconstruction, inverse problems, regularization, total variation regularization, uncertainty quantification, Bayesian calibration, reproducible research, open-source educational software, geoscience education, interactive simulation, non-uniqueness, noise sensitivity
SDG: Quality education
Open Access status: gold
Estimated APC paid: $2,896
Awards: G5193080771, G8899177339
OpenAlex ID: W7225915448
